Superbubble Feedback in Galaxy Formation Ben Keller (McMaster - - PowerPoint PPT Presentation

Superbubble Feedback in Galaxy Formation

Ben Keller (McMaster University) James Wadsley, Samantha Benincasa, Hugh Couchman Paper: astro-ph/1405.2625 (Accepted MNRAS) Keller, Wadsley, Benincasa & Couchman 2014

Background Image: High-resolution simulation of Milky Way like galaxy using superbubble feedback. Outflows with entrained cold clouds can be observed.

Stellar Feedback: Motivation

M82 Image: HST, NASA/ESA

• Feedback from Massive

stars: metals, energy, momentum through Winds, UV, SNII

• FB regulates star

formation, ISM structure FB-driven Galactic winds:

• Remove gas from disk,

enrich IGM with metals

• Set final stellar mass

Superbubble Feedback: Motivation

N70 Superbubble LMC Image: ESO D 100 pc Age: 5 Myr v ~ 70 km/s Driver: OB assoc. 1000+ stars

• Massive star formation

highly correlated in time and space

• Typical star cluster

~ 10,000 Mʘ forms in ~10 pc over < 1 Myr Stellar Feedback highly correlated Natural unit of feedback is a superbubble combining feedback of 100+ massive stars

Super bubble features

MacLow & McCray 1988, Weaver+ 1977, Silich+ 1996

Classic model:

• Stellar winds + supernovae

shock and thermalize in bubble

• Negligible Sedov-phase
• Mechanical Luminosity

L=1034 erg/s/Mʘ

• Much more efficient than

individual SN (e.g. Stinson 2006 Blastwave feedback model )

Super bubble features

Limiting factor: Radiative Cooling of bubble determined by bubble temperature ~ Eth/Mb and density Mb/R3 Hot bubble mass (Mb) set by thermal conduction rate into bubble

MacLow & McCray 1988, Weaver+ 1977, Silich+ 1996

Modeling Superbubbles

• 1. Key physics: Thermal Conduction

Without conduction bubble mass = ejecta mass

• 2. Evaporation resulting from conduction – hard to

resolve directly

• 3. Low resolution, early bubble stages:

Mb < Mparticle – need to avoid overcooling

• 1. Thermal Conductivity
• Self regulating Energy flux ~ T7/2/R (T > 105 K)
• Flux limited by electron speeds (Cowie & McKee

1977)

• Note: κ reduced by 3-5 by Magnetic Fields
• For sharp temperature contrast, drives evaporative

mass flux from cold into hot gas

(cgs) 10 6 ) (

2 / 5 7 T

T t E

Cond Cond 

        

• 2. Evaporation
• Evaporation front width < 0.1

pc ! Subgrid model:

• Based on MacLow & McCray

1988 rate estimate

• SPH implemention:

Stochastically evaporate particles into hot bubble from cold shell

• Applied for T > 105 K particles
• Regulates bubble temperature

2 / 5

25 16 T k t M

b b

    

• 3. Low Resolution : Subgrid Hot Phase
• For a poorly resolved bubble, Mb < Mparticle for

the early stages

• Temporary 2-phase particle while

injection/conduction grows mass of bubble phase

• No numerical/resolution related overcooling
• Feedback-heated particles briefly contain 2

phases in pressure equilibrium, with separate densities and temperatures

– Each cools independently.

• N-body Solver (Tree Method) and

Smoothed Particle Hydrodynamics

• Physics: Gravity, Hydrodynamics, Atomic

Chemistry (Radiative Heating, Cooling), Radiative Transfer (Woods et al, in prep)

• Subgrid Physics: Star Formation,

Turbulent Diffusion

Gasoline

Implementation:

Wadsley+ 2004

High Resolution Superbubble Simulation

Mass loading

• Bubble mass, temperature regulated:

Match Silich+ 1996 Mass loading For 3x1038 erg/s Feedback

Test 30,000 Mʘ cluster: 3 cases

Direct Injection: Resolved stellar ejecta mass, no subgrid required (Mparticle=760 Mʘ at 1283), conduction + evaporation Superbubble: conduction, evaporation + subgrid Simple Feedback: A non-cooling phase with conversion time 5 Myr to cooling form (cf. Agertz+ 2013)

Keller+ 2014

Bubble Momentum + Hot Mass

• Simple Model resolution sensitive
• Superbubble Model still works with a 1 particle bubble

(323 case)

Keller+ 2014

Time (Myr) Hot Bubble Mass Bubble Momentum

Galaxy Tests

Similar to Dalla Vecchia & Schaye (2012) -- MW analogue (Mgas ~ 109 Mʘ Ngas = 105) & Dwarf

Keller+ 2014

MW Analogue: Temperature & Column Density

MW & Dwarf Star Formation

Keller+ 2014

• Star formation rates
• regulated. Bursty as

expected in dwarf

• Higher mass loading
• Outflow evolution

similar to Dalla Vecchia & Schaye 2012

• Note: dwarf has low

surface density

• Kennicutt-Schmidt law

matched

Galaxies: SFR & Outflows

Milky Way Dwarf

Keller+ 2014

Outflow Rate & SFR Outflow Velocity Time (Myr) Time (Myr)

Temperature-Density Phase space

No gas in short cooling time region Particles split into cold dense + hot rarefied phases Rapidly become hot, single phase – evolve adiabatically

Keller+ 2014

Summary

• Superbubble is relevant scale for stellar feedback in galaxies
• Thermal conduction is dominant physical process in

superbubble evolution

• Taking this into account gives you a powerful model for

feedback:

– Separating Cold & Hot phases in unresolved superbubble prevents

• vercooling

– Feedback can be continuous, multi-source – Feedback gas doesn’t persist in unphysical phases – Star formation is strongly regulated, winds are driven with realistic mass loadings

• Read the Paper:

– astro-ph/1405.2625 (Accepted MNRAS) – Keller, Wadsley, Benincasa & Couchman 2014

Stellar Feedback Budget

Starburst ‘99 Erg per Mʘ

Time (years) Bolometric Luminosity Supernovae Type II Winds UV Energy Injection Rate (log10 erg/s/Mʘ)

• UV & Radiation

pressure disrupt dense clouds

– Denser gas (>104 H/cc) dispersed, star formation cut off

• SNII and stellar winds

Steady 1034 erg/s/Mʘ

for ~ 40 Myr

Super bubbles: Vishniac Instabilities

Theory: Vishniac 1983 Sims: McLeod & Whitworth 2013, Nayakshin+ 2012 (AGN) Nirvana simulations 3 star bubble Krause et al 2013

Super bubbles: X-Ray Observations

Chu 2008

Krause+ 2014

• X-Ray luminosity highly

variable over space, time

• Very few observations,

large scatter in observed LX

• Leaking of interior, B-

field amplification in shell may explain some reduced luminosities (see Rosen+ 2014)

Clumpy medium

Clumpy Medium

• Some changes in bubble mass/momentum
• Agreement with direct model still good

Reduced Conduction & Magnetic Fields

• Conduction

suppressed across magnetic field lines

• 100x reduction in

conduction rate κ0 results in only factor of ~2 reduction in Mb

Multiphase Properties

• Median time as mixed-phase particle < 5 Myr

Cosmological Galaxy (now z=2)

• ~ 10 11 Msun halo
• So far on track for reasonable M *

Coming Soon…